DENV3 genotype I (MG20), which is highly virulent for mice, was obtained from virus collection of Laboratório de Vírus of Universidade Federal de Minas Gerais (UFMG), Belo Horizonte, MG, Brazil. It was isolated from a patient presenting neurological manifestations of dengue and death [23] . The virus used in this study was a virus with only 6 passages in C6/36 cells, thus maintaining the initial highly virulent characteristics of the virus.

C6/36 cells is a continuous lineage obtained from Aedes albopictus (ATCC, number CRL-1660) larvae and were cultivated on Leibovitz (L-15) medium (Gibco, USA), containing ciprofloxacin at 10 μg/mL and fetal bovine serum (FBS) at 10% (Cultilab, Brazil). Cells were maintained in BOD at 28˚C. For the passage of cell monolayers, culture medium was discarded and cells washed with phosphate buffer saline (PBS), pH 7,2 and then homogenized in L15 medium (Gibco, USA) and distributed into 25 or 75 cm² culture flasks. C6/36 cells were used to study viral replication. BHK-21 cells derived from hamster kidney (ATCC, number CCL-10) were cultivated on Dulbecco’s Modified Eagle Medium (DMEM, Gibco, USA), containing gentamicin at 50 μg/mL, penicillin at 100 IU/mL, amphotericin B at 5 μg/mL and FBS at 5%. Cells were maintained in a CO₂ cell incubator at 37˚C. BHK cells were used in viral titration protocol and serum neutralization test.

Immune competent C57BL/6 mice were provided by animal facilities of Instituto René Rachou-Fiocruz-MG (IRR). The animals, all males, were 8 weeks old.

This project was submitted to and approved by the Ethical Committee of Animal Handling (CEUA) from Fundação Oswaldo Cruz, Fiocruz, Brazil (LW-7/16), see annex 1. This study was carried out in strict accordance with the recommendations of the Brazilian National Council for Control of Animal Research (CONCEA).

In previous publication we demonstrated that mice inoculated by intracranial route with highly virulent virus presented clinical signs of disease as loss of weight, lethargy, ruffled fur, hunched posture and paralysis of paws, and since then it has been established by our group that the mice with these clinical signs should be immediately euthanized [6] . These criteria were described in the project submitted and approved by the Ethical Committee of Animal Handling described above.

Additionally the infected or mock mice were euthanized at the end of each experiment. No animal died before meeting the euthanasia criteria or before the end of the experiment.

Animals were maintained in micro isolators on ventilated shelves at IRR’s animal facilities, where the temperature (23˚C ± 2˚C), humidity, ventilation, sanitation and illumination were controlled. Food and water were provided ad libitum. Mice received subcutaneous inoculum (SC) containing 5 × 10⁴ p.f.u. of DENV 3 genotype I (MG20) on a final volume of 100 μL. The animals were monitored daily for evaluation of clinical signs and survival curve and weighed daily. All cages had igloos as environmental enrichment for animals. All those involved in the handling of the animals were previously trained by the IRR’s animal facilities.

Subcutaneous infection (SC)―the mice received inoculum of 5 × 10⁴ p.f.u. of DENV by SC route (inoculum of 100 uL); Intracranial infection (IC)―the mice received inoculum of 400 p.f.u. of DENV by IC route (inoculum of 10uL); Subcutaneous infection followed by intracranial infection (SC-IC)―the mice received 5 × 10⁴ p.f.u. of virus (inoculum of 100 uL) and after 14 days received new infection by IC route with 400 p.f.u. of virus (inoculum of 10 uL); Mock mice―the mice received only inoculum with the C6/36 cell supernatant. Six mice were used in each group.

BHK-21 cells were used to titer viral stocks and the organs of infected mice. Cell monolayer, 5 × 10⁵ BHK-21 cells were implanted in 6 wells plates with DMEM supplemented with 5% of heat inactivated FBS (Cultilab, Brazil) and antibiotics, and incubated for 24 hours at 37˚C in a 5% CO₂ cell incubator. Medium was discarded and cell monolayer was washed once with PBS. Then 400 μL of serial dilutions (using 10⁻³ to 10⁻⁷ dilutions) of virus to be titrated were added in each well, leaving a well for cell control. After one hour of adsorption the medium was removed and 2 mL/well of DMEM, containing 1% of carboxymethylcellulose (CMC) (Sigma, Germany) and 2% of FBS was added. Cells were incubated at 37˚C for 6 days at a cell incubator with 5% of CO₂. After 6 days of observation under a microscope, cells were fixed for 1 hour with formaldehyde at 3.7% diluted in PBS for subsequent staining with crystal violet at 1% solution for 20 minutes. The number of lysed plaques were counted for titer determination in plaque forming units per milliliter (p.f.u./mL).

The adapted protocol from Russell et al. [24] was used for serum neutralization test. BHK-21 cells were plated at the concentration of 5 × 10⁴ cells/ml in 24 well plate (1 mL/well) with DMEM supplemented with 5% of FBS and antibiotics and incubated for 24 hours at 37˚C in a 5% CO₂ cell incubator. Sera were inactivated at 56˚C for 30 minutes and diluted (1:10, 1:20, 1:40: 1:80) in DMEM supplemented with 1% FBS and incubated with the same volume of diluted virus (50 p.f.u./well) for 1 hour at room temperature. The neutralized sera were added to BHK-21 cells in plates for one hour for adsorption. The serum was removed and 1 mL DMEM supplemented with 5% of FBS and 1% CMC was added in each well and was incubated at 37˚C for 7 days in a 5% CO₂ cell incubator. After seven days the cells were fixed for one hour with 3.7% formaldehyde diluted in PBS and subsequently stained with crystal violet (1%) for 20 minutes. Plaques were counted in wells that received negative serum and, in the wells, which received the test sera. The PRNT₅₀ index was calculated by the highest serum dilution capable of reducing the total number of plaques by 50%, relative to the control.

To measure the cytokine levels in brain tissues were collected, macerated and stored on a protease inhibitor solution (complete, Mini, EDTA-free) and 25 μL/well of organ macerate or 25 μL of each dilution of standard (including the negative control) were added into a 96 well V-bottom plate. Following, 18 μL/well of the CBA mix (Inflammation BDTM Cytometric Bead Array (CBA) mouse kit [25], USA) was pipetted. Then, 15 μL of PE detection reagent was added and it was incubated for 2 hours protected from light. After the incubation period, 100 μL of wash buffer was added to each well and centrifuged at 200 g for 5 minutes. Supernatant was discarded and 150 μL/well of wash buffer was added to wash the beads. Samples were read in the FACSCan (BD Biosciences) and the analysis of the CBA was done in the CellQuest software (BD Biosciences), with the results represented in pg/mL. The cytokine kinetics was chosen based on innate (6 and 12 h.p.i. hours post infection) and adaptive immune response (8 d.p.i. days post infection) in the host [19] and on the time at which the weight loss begins.

Real-Time PCR was performed to measure levels of virus genomic RNA expression in brain and liver [26] [27] . RNA extracted with TRIzol® reagent was treated with DNAse prior to further reverse transcription. Protocol of RNA treatment with DNAse was according to the RQ1 RNAse-Free DNAse kit instructions (Promega). Procedure for reverse transcription (RT) was performed according to the protocol of the manufacturer (Promega Corporation, USA). For the qPCR, 2.5 μL of the cDNA obtained from the RT were used. Each primer (0.2 μL) was aliquoted at 10 μM; 5.0 μL of Power SYBR® Green PCR Master Mix 2X (Applied Biosystems) and 2.0 μL of RNAse-free water (Ultra-Pure Distilled Water-Invitrogen) with a final volume of 10.0 μL/well was used in a 384 wells plate containing SYBR Green I marker, AmpliTaq Gold DNA Polymerase (2 units for reaction), dNTP, buffer and passive reference dyestuff ROX, in Applied Biosystems ViiA™ 7 apparatus. Amplifications were used with 40 cycles of 95˚C for 30 seconds and 60˚C for 1 minute, preceded by 95˚C for 3 minutes. After the reaction, the amplified specificity was observed, conferring the dissociation temperature (TM) given by the dissociation curve, specific for the target transcript. The following oligonucleotides were used in the reactions: HPRT―hypoxanthine guanine phosphoribosyl transferase (Reverse: 5’-GAT TCA ACT TGC GCT CAT CTT AGG-3’; Forward: 5’-GTT GGA TAC AGG CCA GAC TTT GTT G-3’); DENV-5’ URT DENV Non-coding 5’ terminal region of the DENV: (Reverse: 5’-TCC GTT GGT TGT TCA TCA GA-3’; Forward: 5’-TCG GAA GCT TGC TTA ACG TAG-3’). Relative quantification methodology was used to analyze the data. Expression of the genomic RNA was normalized to the expression level of the constitutively expressed HPRT gene.

Immediately after anesthesia, blood samples were obtained through brachial plexus and were added on microtainer tubes coated with EDTA (Becton Dickinson Vacutainer Systems, Franklin Lakes, NJ, USA). Red blood cells (diluted in PBS 1:200), platelets (diluted in ammonium oxalate 1:100) and leukocytes (diluted in Turk’s blue solution 1:20) counts were performed manually on a Neubauer chamber and monocytes, neutrophils and lymphocytes were counted manually on a previously fixed and stained blood smear slide.

The infected groups were compared to the control group using the Mann-Whitney test for non-parametric samples and T-test for parametric samples. Statistical analyzes were performed utilizing the statistical software GraphPad Prism 5 (GraphPad Software, Inc., La Jolla, CA). Differences were considered significant when p < 0.05.

The C57BL/6 mice were inoculated via intracranial (IC) route with 4 × 10² p.f.u. of DENV 3 genotype I (Figure 1). Clinical signs (weight loss, pilo-erection, hunched posture and hind limb paralysis) were evaluated daily, confirming the susceptibility to neuro-infection of mice by via intracranial (IC) route which developed neuro-encephalitis and death within 9 days after infection, as reported previously [6] . In another model, the same highly virulent DENV and the same strain of mice used in previous studies but inoculated subcutaneously (SC)

instead IC. This subcutaneously (SC) infection was performed with 5 × 10⁴ p.f.u. and the mice were evaluated daily for the appearance of clinical signs of the disease during the next 25 days (Figure 1).

The infected mice presented a percentage variation of weight gain lower than the uninfected ones. In this model of infection, the animals did not show pilo-erection, hunched posture or hind limb paralysis, and no mouse died, but there was weight loss in infected mice compared to uninfected mice. The infected mice presented this weight loss from day 10 post infection and regained weight from day 18 after infection, maintaining that gain until the end of the experiment.

Firstly, a SC infection was performed with 5 × 10⁴ p.f.u. and on day 14 after this infection, a second infection, now via IC, was performed with 4 × 10² p.f.u. of highly virulent DENV. The variation of weight gain between the infected and the non-infected mice were similar to that in the SC infection and mock SC only. The mock curve was performed with mean of the weights of the mice inoculated separately with C6/36 cell supernatants by the three routes (SC, IC and SC-IC), because the weights were very similar.

To confirm virus effective infection via IC, SC and SC-IC, virus brain titration, with inoculum via IC or via SC or via SC-IC or serum neutralization via inoculum SC, of infected mice, were performed (Table 1) and also Real Time PCR was performed to confirm the presence of genomic RNA in the brain (infected via IC or SC-IC) or in the liver of the animals infected by SC (Figure 2).

The results confirm that the virus replicated inside the brain in the IC or SC-IC or in the serum via SC infection with a titer of 1.2 to 1.4 × 10⁴ p.f.u. at the 5^th day post infection and 29 × 10⁴ p.f.u. at the seventh day post infection (Table 1). There was an antibody response against DENV in the SC infection at a titer of 1:80 per mL at the 60^th day post infection (Table 1).

Leukopenia was observed in subcutaneously infected mice on the 8^th day after infection when compared to uninfected animals (Figure 3(a)). There was no significant difference in erythrocytes counts in mice (Figure 3(b)) in any of the times studied when compared to the uninfected ones. Figure 3(c) shows a decrease in the number of platelets at all times studied in the infected mice when compared to the uninfected mice.

For virus titration in the brain after IC inoculation with 400 p.f.u. of DENV or with C6/36 cell supernatant (mock), C57BL/6 mice were euthanized for removal of the brain on days 3, 5 and 7 post-infection. For the titration of the virus in the brain after SC-IC infection, C57BL/6 mice were inoculated via SC with 5 × 10⁴ p.f.u. of DENV or with C6/36 cell supernatant (mock) and 14 days after this infection, the mice were inoculated again via IC with 400 p.f.u. of DENV or with C6/36 cell supernatant (mock) and mice were euthanized for removal of the brain on days 3, 5 and 7 after this second inoculation. For virus titration in the serum, after inoculation via SC with 5 × 10⁴ p.f.u. of DENV or with C6/36 cell supernatant (mock), the serum was collected 3, 5 and 7 days after inoculation. For the seroneutralization test (PRNT50), C57BL/6 mice were inoculated via SC with 5 × 10⁴ p.f.u. of virus or with C6/36 cell supernatant (mock) and serum samples were collected 60 days after inoculation. d.p.i. = days after infection.

There was no difference in lymphocytes counts in infected mice, when compared to noninfected mice (Figure 4(a)). A decrease in the monocytes count was observed 6 hours after infection in mice, when compared to non-infected mice (Figure 4(b)). In the infected mice there was no difference in the neutrophils count in comparison to the uninfected mice, at the times studied (Figure 4(c)).

An increase of TNF alpha level in mice spleen, 6 hours after subcutaneous DENV inoculation was observed, when compared to the uninfected group (Figure 5(a)). An increase of TNF alpha level in mice liver was observed 6 and 12 hours after subcutaneous DENV infection, when compared to the uninfected group (Figure 5(b)). The cytokine IL-12p70 in the liver of infected mice was similar to cytokine level found in uninfected mice at all times studied (Figure 5(c)). On the other side, IFN gamma in the liver of mice 6 and 12 hours after infection with DENV was increased, when compared to the uninfected mice (Figure 5(d)). In the spleen no IL12-p70 and IFN gamma were detected (data not shown).